Influenza pandemics occur when an abrupt change in the composition of the circulating viruses results in a strain to which individuals have little or no immunity. The three pandemic outbreaks that occurred in the 20th century in 1918, 1957, and 1968 represent three different antigenic subtypes of influenza A virus: H1N1, H2N2, and H3N2, respectively. The concern of a future pandemic has focused attention on the avian H5N1 virus, which has demonstrated not only the possibility of infecting humans but to do so with a high case-fatality rate. Vaccinating populations to protect them from a future pandemic virus poses an obvious problem. David Nabarro, UN senior coordinator for avian and human influenza, stated: "We cannot have a vaccine against a pandemic virus until the pandemic virus appears and then there will be an interval of perhaps six months until we have a reasonable quantity of vaccine." Efforts are underway to create pre-pandemic vaccines to H5N1 viruses that can confer broad protection against different strains. However, the divergence of highly pathogenic H5N1 viruses into distinct clades has increased the complexity of developing a prophylactic vaccine against multiple H5N1 subtypes. We propose here a novel way to approach this problem, using the power of directed molecular evolution. This technology encompasses a method for creating novel protein sequences coupled with a means of screening the protein variants. In vitro homologous DNA recombination can create libraries of high-quality functional diversity from pathogen genes encoding the proteins needed to create a vaccine. Libraries of the variant proteins can then be screened for improved immunogenicity. The only major correlate of protection against influenza is antibody to its surface glycoproteins, primarily the hemagglutinin (HA). We hypothesize that variants of the H5 HA protein can be created that are more immunogenic than the wild-type proteins, and will induce immune sera with potentially broad neutralization potency. Genes encoding the variants to test this hypothesis will be created by in vitro DNA recombination of wildtype H5 HA-encoding genes. The specific product that is being targeted by this work is an HA-based vaccine with broad neutralization activity against many H5N1 viruses. The creation of a stockpile of variants and the sera induced by them will nonetheless provide a means to rapidly choose the immunogen needed to prepare a vaccine once a pandemic strain has appeared. The current proposal will assess the feasibility of this idea. A study by the Congressional Budget Office estimates that the consequences of a severe pandemic could include 200 million people infected, 90 million clinically ill, and 2 million dead in the United States alone. If a pandemic is caused by H5N1, the numbers of deaths could be considerably higher given the observed >50% case-fatality rate for infection by this virus. A strategy such as that proposed here could - if successful - provide a measure of protection against a pandemic H5 influenza strain and potentially save millions of lives. PUBLIC HEALTH RELEVANCE: Influenza pandemics occur when individuals are poorly protected from the virus. We propose to create a collection of vaccine candidates that can rapidly be tested against an emerging pandemic. A strategy such as that proposed here could provide protection against a pandemic flu virus and save millions of lives.